Medication Prescriptions in 322 Motor Functional Neurological Disorder

General Hospital Psychiatry 58 (2019) 94–102

Contents lists available at ScienceDirect

General Hospital Psychiatry

journal homepage: www.elsevier.com/locate/genhospsych

Medication prescriptions in 322 motor functional neurological disorder patients in a large UK mental health service: A case control study T ⁎ Nicola O'Connella, , Timothy Nicholsona, Graham Blackmana, Jennifer Tavenera, Anthony S. Davidb a Institute of Psychiatry, Psychology and Neuroscience, King's College London, 16 DeCrespigny Park, London SE5 8AF, United Kingdom b Institute of Mental Health, UCL Division of Psychiatry, University College London, 6th Floor, Maple House, 149 Tottenham Court Road, London W1T 7BN, United Kingdom

ARTICLE INFO ABSTRACT

Keywords: Objective: This study describes medication prescribing patterns in patients with motor functional neurological Functional neurological disorder disorder (mFND) treated in South London and Maudsley NHS Foundation Trust (SLaM), comparing outcomes to Functional motor disorder a control group of psychiatric patients from the same hospital trust. Psychosomatics Method: This is a retrospective case-control study using a psychiatric case register. Cross-sectional data were Conversion disorder obtained from 322 mFND patients and 644 psychiatry controls who had had contact with SLaM between 1st Polypharmacy January 2006 and 31st December 2016. Results: A slightly lower proportion of mFND patients received medication compared to controls (76.6% v. 83.4%, OR: 0.59, CI: 0.39–0.89, p < 0.05). Of medication recipients, mFND patients were prescribed a higher number of agents (mean: 4.7 v 2.9, p = 0.001) and had higher prescription rates of antidepressants, anti-epileptics, analgesics, and certain non-psychotropic medications. Higher numbers of prescriptions were associated with co-morbid physical conditions, and previous psychiatric admissions. Conclusions: This is the first study to describe medication prescriptions in a large cohort of mFND patients. Patients were prescribed a wide range of psychiatric and physical health medications, with higher rates of polypharmacy than controls. Psychotropic medication prescription is not necessarily the first line treatment for mFND, where physiotherapy and psychotherapy may be offered initially. There is limited, early-phase evidence for pharmacological therapies for mFND, and as such, the benefit-to-risk ratio of prescribing in this complex and poorly understood disorder should be carefully assessed.

1. Introduction the seizure variant of FND, and specialist physiotherapy for motor FND (mFND) [6]. There have been no large randomised controlled trials In functional neurological disorder (FND), also known as conversion (RCTs) of pharmacological treatments and much of the evidence on disorder, patients present with neurological symptoms with no known pharmacotherapy comes from extrapolations from other functional neurological cause [1]. Symptoms commonly include seizures, sensory disorders such as irritable bowel syndrome (IBS), fibromyalgia and (e.g. numbness or visual impairment) and motor symptoms such as multiple functional disorders [1]. A recent FND review states that it is weakness, movement (e.g. tremor, paralysis and dystonia) and gait inappropriate to prescribe medication for functional symptoms speci- disorders. FND is common and associated with high rates of disability fically, although comorbidities like pain and depression may be treated [2], lower quality of life [3,4] and poor prognosis [5]. FND has, until pharmacologically [6]. recently, fallen between the disciplines of Neurology and Psychiatry Of existing pharmacotherapy research in mFND, a small open-label and has been under-researched. Consequently, there is a poor evidence study reported some effectiveness for the antidepressants citalopram base for treatment. and paroxetine [7], although all patients who improved had co-morbid There is emerging evidence for the effectiveness of cognitive be- anxiety or depression. A small, unblinded observational study of 18 havioural therapy (CBT) for psychogenic non-epileptic seizures (PNES), functional patients, of whom nine had paralysis and two had ataxia,

⁎ Corresponding author. E-mail addresses: nicola.k.o'[email protected], [email protected] (N. O'Connell), [email protected] (T. Nicholson), [email protected] (G. Blackman), [email protected] (J. Tavener), [email protected] (A.S. David). https://doi.org/10.1016/j.genhosppsych.2019.04.004 Received 1 March 2019; Received in revised form 1 April 2019; Accepted 2 April 2019 0163-8343/ © 2019 Elsevier Inc. All rights reserved. N. O'Connell, et al. General Hospital Psychiatry 58 (2019) 94–102 reported greater improvement in patients treated with sulpiride com- epileptic medications in mFND. pared to haloperidol, although patients may have improved regardless of drug intervention [8]. There were beneficial responses to both ben- 2. Methods zodiazepines and placebo in a study of 26 patients with ‘psychogenic paroxysmal movement’ disorder [9]. In the seizure variant of FND, a 2.1. Design and source of clinical data pilot double-blind trial found sertraline reduced the frequency of nonepileptic seizures [10] and an open-label study of patients with PNES This was a retrospective case-control study which includes patients and anxious and depressive symptoms treated with venlafaxine also in contact with South London and Maudsley (SLaM) Foundation Trust, reported seizure reductions and improvements in co-morbid anxiety between 1st January 2006 and 31st December 2016. and depression [11]. Data were obtained from the SLaM Biomedical Research Centre's In other functional disorders, pharmacological guidelines differ but (BRC) Clinical Records Interactive Search (CRIS) database which con- antidepressant use is common [12]. Low-dose imipramine has been tains anonymised electronic health records of over 250,000 individuals found to improve overall health in patients with multiple functional referred to SLaM services [38]. SLaM provides inpatient and commu- symptoms compared to placebo [13], and there is low-quality evidence nity services for a catchment of 1.5 million people living in southeast for the efficacy of new-generation antidepressants in unexplained London. The trust also receives national referrals for FND patients. physical symptoms [14]. Within specific disorders, antidepressants We retrieved medication information from searches of the databa- have shown improvements in IBS [15,16], chronic pain [17] and so- se's free clinical notes fields. As some patients had contact with SLaM matoform pain disorder [18], with some evidence for their efficacy in over the course of years, the most contemporaneous medication in- patients without concomitant mood disorders [13,19]. Most studies do formation was collected from each participant, representing a cross- not screen for comorbidities however, and the mechanisms of anti- sectional snapshot at one timepoint, rather than a cumulative retro- depressant effectiveness are poorly understood [20]. spective list of medications. In some cases, no information was available Medication prescriptions rates are high across functional disorders. on medication at all. This was marked, ‘not known’. Given our data High prescription rates have been reported in patients with ‘psycho- collection method and the necessary trade-off between capturing rich genic jerky movement disorder’ and in patients with unexplained pain, clinical information on a wide range of variables and increasing our compared to Tourette's patients and those with pain disorders respec- sample size, we capped the number of medications we recorded per tively [21,22]. Opioid prescriptions varied between 28 and 36% in patient at 12. Strictly, this should be interpreted as “12 or more med- patients with unexplained gastrointestinal symptoms, fibromyalgia, and ications”. We calculated the total number of medications each partici- chronic fatigue [23–25] and higher rates of benzodiazepines, gastric pant was prescribed. We further categorised each drug based on their reflux medications, inhalers and blood pressure medications have been therapeutic properties, guided by the NICE British National Formulary reported in PNES compared to epilepsy patients [26,27]. Patients with (BNF) [39]. Our classification led to a total of 27 drug categories. functional symptoms in primary care have been reported to have higher rates of both psychotropic and somatic medications compared to con- 2.2. Study setting and participants trols [28]. The lack of pharmacotherapy evidence in mFND means prescribing Cases were defined as any patient aged over-18 years with a primary patterns will vary and may be more likely to be infl uenced by a range of or secondary diagnosis of ‘Conversion disorder with motor symptom or individual and systemic factors than prescriptions in non-FND patients. deficit’ (ICD-10 code: F44.4) who had contact with SLaM services. In While psychiatrists will prescribe for common mFND comorbidities addition, patients with any F44 diagnosis were included if there was such as insomnia, pain, depression, anxiety and panic [29,30], clin- evidence of functional motor symptoms in their unstructured case notes icians may also be influenced by patients' personal characteristics [31], or correspondence (see “Supplementary materials” for a comprehensive symptom severity and treatment preferences [12]. Some GPs have de- list of search strategies). scribed using prescriptions for patients with unexplained symptoms on Our control group comprised contemporaneous SLaM patients who a trial-and-error basis [32]. The overall threshold for prescribing may received any non-functional (i.e. non-F44) psychiatric diagnosis on the be lower as a result of the fewer treatment options [33]. A lack of succeeding day the mFND patient received their diagnosis. Patients evidence may also lead some clinicians to prescribe placebos [34], with were included if they were aged over-18 years and had contact with attendant complex ethics [35]. On a systemic level, functional patients SLaM services. Those with a neurodegenerative or intellectual dis- have increased contact with and referrals within mental and physical abilities (F70–F79) diagnosis were excluded. We used a random number health services [36,37], which may lead to greater numbers of pre- generator from the website, random.org to select controls from the scriptions and a cycle of medicalization may occur where patients re- returned search results and adopted a case-control ratio of 1:2. quire additional treatment to manage the side effects of a primary medication, for example requiring laxatives due to opioid analgesic 2.3. Ethical approval prescriptions. Given the potential for iatrogenic harm, there is a conspicuous lack CRIS has received ethical approval from the Oxfordshire Research of evidence on the type and number of medications mFND patients Ethics Committee C (08/H0606/71+5) as an anonymised dataset for receive. mental health research. Ethical approval was granted in 2008 and re- The aim of this study was to describe cross-sectional medication use newed for a further five years in 2013. This study was approved by a in patients with mFND treated in a large secondary care mental health patient-led NIHR BRC CRIS oversight committee (CRIS 14-101). Trust. Our objectives were to: 2.4. Statistical analysis i) To test the hypothesis that patients with mFND are prescribed more medications and across a larger range of classes than a random Descriptive statistics were used to describe socio-demographics and sample of contemporaneous psychiatric patients treated in the same patients' prescriptions. Means, standard deviations, medians and in- hospital trust; terquartile ranges (IQR) were used to describe continuous data and ii) explore the socio-demographic variables and symptoms linked to proportions to define categorical data. We employed odds ratios (OR) polypharmacy in mFND; and with 95% confidence intervals to compare the risk of medication pre- iii) examine the socio-demographic and symptom factors linked to scription in both groups. prescriptions of opioid analgesics, antidepressants and anti- Univariate associations between prescription of five or more

95 N. O'Connell, et al. General Hospital Psychiatry 58 (2019) 94–102

Table 1 Use of prescription medications in mFND and control group patients.

Number and proportion of patients prescribed one or more medicationa Total prescriptionsb

mFND Control OR 95% CI p value mFND Control Chi2 95% CI p value n (%) n (%) n (%) n (%)

Alcohol/drug treatment 7 (2.8) 37 (6.9) 0.4 0.2–0.9 0.03 7 (0.6) 40 (2.5) 14.6 0.9–2.8 0.001 Antidepressant 168 (68) 283 (52.7) 1.9 1.4–2.6 0.001 196 (16.8) 310 (19.3) 2.8 −0.4–5.40 0.09 Anti-epileptic 84 (34) 84 (15.6) 2.8 1.9–3.9 0.001 98 (8.4) 91 (5.7) 7.7 0.8–4.7 0.005 Anti-histaminec 27 (10.9) 33 (6.1) 1.9 1.1–3.2 0.02 27 (2.3) 33 (2.1) 0.28 −0.8–1.5 0.59 Sedating 4 (14.8) 21 (70) 16.5 29.5–71.3 0.001 Non-sedating 21 (77.8) 9 (30) Anti-inflammatory 24 (9.7) 9 (1.7) 6.3 2.9–13.8 0.001 24 (2.1) 9 (0.6) 12.5 0.6–2.5 0.001 Anti-manic 4 (1.6) 15 (2.8) 0.5 0.2–1.7 0.45 4 (0.3) 15 (0.9) 3.8 −0.02–1.2 0.052 Anti-microbial 18 (7.3) 20 (3.7) 2.0 1.1–3.9 0.03 22 (1.9) 24 (1.5) 0.7 −0.6–1.5 0.42 Anti-muscarinic 18 (7.3) 26 (4.8) 1.5 0.8–2.8 0.17 18 (1.5) 27 (1.7) 0.2 −0.8–1.1 0.68 Anti-nausea 15 (6.1) 11 (2.0) 3.1 1.4–6.8 0.005 16 (1.0) 12 (0.7) 0.7 −0.4–1.1 0.39 Antipsychotic 29 (11.7) 236 (43.9) 0.16 0.1–0.3 0.001 30 (2.6) 265 (16.5) 137 11.9–15.9 0.001 Anti-asthma 31 (12.6) 30 (5.6) 2.4 1.4–4.1 0.001 41 (3.5) 38 (2.4) 2.9 −0.2–2.5 0.09 Benzodiazepines 41 (16.6) 71 (13.2) 1.3 0.9–1.9 0.21 43 (3.7) 76 (4.7) 1.6 −0.6–2.5 0.19 Beta blocker 13 (5.3) 29 (5.4) 0.9 0.5–1.9 0.94 13 (1.1) 30 (1.9) 2.8 −0.2–1.7 0.09 Cardiac/hypertension 42 (17) 52 (9.7) 1.9 1.2–2.9 0.004 64 (5.5) 78 (4.9) 0.5 −1.0–2.3 0.48 CNS Stimulant 2 (0.8) 17 (3.2) 0.25 0.1–1.1 0.06 2 (0.2) 22 (1.4) 10.9 0.5–1.9 0.001 Corticosteroid 24 (9.7) 22 (4.1) 2.5 1.4–4.6 0.003 28 (2.5) 26 (1.6) 2.8 −0.2–2.1 0.09 Diabetes medication and insulin 15 (6.1) 22 (4.1) 1.5 0.8–2.9 0.23 18 (1.5) 30 (1.9) 0.6 −0.6–1.4 0.43 therapy Hormone replacement 30 (12.1) 26 (4.8) 2.7 1.6–4.7 0.001 35 (3) 27 (1.7) 5.2 0.2–2.6 0.02 Laxatives/bowel dysfunction 50 (20.2) 32 (6.0) 4.0 2.5–6.4 0.001 70 (6.0) 38 (2.4) 23.3 2–5.2 0.001 Nicotine replacement 3 (1.2) 5 (0.9) 1.3 0.3–5.5 0.71 5 (0.4) 5 (0.3) 0.2 −0.3–0.7 0.65 Nutrient 49 (19.8) 83 (15.5) 1.4 0.9–2.0 0.13 71 (6.1) 146 (9.1) 8.4 0.9–4.9 0.004 Opioid analgesic 77 (31.2) 28 (5.2) 8.3 5.2–13.1 0.001 89 (7.6) 35 (2.2) 46.1 3.8–7.1 0.001 Other 21 (8.6) 38 (7.2) 1.2 0.7–2.1 0.52 21 (1.8) 38 (2.4) 1.2 −0.5–1.6 0.28 Painkiller (non-opioid) 81 (32.8) 48 (8.9) 4.9 3.3–7.4 0.001 99 (8.3) 52 (3.2) 34.7 3.4–6.9 0.001 Proton pump inhibitors 58 (23.5) 48 (8.9) 3.1 2–4.7 0.001 58 (4.9) 49 (3.1) 5.9 0.3–3.4 0.02 Statin 35 (14.2) 41 (7.6) 1.9 1.2–3.2 0.005 36 (3.1) 43 (2.7) 0.4 −0.8–1.7 0.53 Muscle relaxant 15 (6.1) 3 (0.6) 11.5 3.3–40.1 0.001 15 (1.3) 3 (0.2) 12.4 0.5–1.9 0.001 Hypnotic 16 (5) 41 (6.4) 0.84 0.5–1.5 0.56 16 (1.4) 42 (2.6) 4.7 0.1–2.2 0.03 Total 1166 (100) 1604 (100) p values in bold denote statistically significant results. a Number and proportion of patients prescribed one or more medication: number and (%) of patients who received one or more of each drug class in both mFND and control groups. Denominator is the number of patients receiving medication. Odds ratio (OR) refers to risk of mFND receiving one or more prescriptions of that drug class compared to control participants. b Total prescriptions: number and (%) of all prescribed drugs according to drug class in both groups. Denominator is the total number of prescribed medications in each group. Chi2 refers to the differences in frequencies of total prescribed medications. c Anti-histamines further classified according to whether they are sedating or non-sedating. 5 medications did not have enough information to classify. medications in mFND patients and patients' socio-demographics and in mFND was ‘Neurotic, stress and somatoform disorders’ of which symptoms were examined with “prescription of one to four medica- ‘Dissociative and Conversion Disorders’ (F44) was the most common. In tions” as the base outcome. Adjusting for all demographic and symp- total 6.8% of mFND patients had a primary mood disorder diagnosis toms factors, a binary logistic regression model was conducted. We and the most common type was a single episode of major depression chose ‘five or more medications’ as five was the median number of (F32). The most frequent secondary diagnosis, after neurotic stress and medications in the mFND group and is a commonly applied definition somatoform disorders were mood disorders, affecting 5.4% of the of ‘polypharmacy’ [40,41]. mFND group (see Table 4, in ‘Supplementary materials’). We conducted a univariate analysis of associations between socio- mFND patients were most commonly given their diagnosis in psy- demographic and symptoms and prescriptions of opioid analgesic, an- chiatric and general hospital outpatient clinics (39.9%), neu- tidepressant and anti-epileptic medications in mFND patients with “no ropsychiatry liaison services in general hospital inpatient settings prescriptions of these drugs” as the base outcome. In a binary logistic (17.7%) and liaison psychiatry in inpatient settings (12.6%). regression model, we adjusted for significant univariate socio-demo- The control group comprised patients with primary diagnoses of graphic variables and examined the associations between mFND mood disorders (22.7%), mental and behavioural disorders due to symptoms and prescriptions of each drug. psychoactive substances (17.4%), and schizophrenia, schizotypal and SPSS for Windows (SPSS v21.0, Chicago, Illinois, USA), Microsoft delusional disorders (14%). For a full description of control group di- Excel (Microsoft Office Professional Plus 2010, Version agnoses see O'Connell et al. [42]. Control participants received their 14.0.7015.1000) and GraphPad Prism (Version 5.01, GraphPad diagnosis in liaison psychiatry services in inpatient settings (20.2%), A Software, La Jolla California USA) were used to analyse data and create &E (14.9%), drug and alcohol services (11%) and assessment and figures. liaison neighbourhood teams (11%). The most commonly reported symptom was ‘weakness’ of any type 3. Results accounting for 50.3% of all reported symptoms, followed by ‘other’ motor or sensory symptoms (37.9%) such as visual disturbances, facial Our search returned 322 mFND and 644 control patients. Socio- droop, etc., and ‘tremor’ which includes ‘tremor, spasms, jerks and tics’ demographic characteristics of both groups are described in Table 4 (33.9%). (see ‘Supplementary materials’). The most frequent primary diagnosis

96 N. O'Connell, et al. General Hospital Psychiatry 58 (2019) 94–102

3.1. Medication prescriptions between polypharmacy and having a co-morbid physical health condition (OR: 6.1, 95% CI: 1.8–19.8, p = 0.003), a previous psychiatry A lower proportion of mFND patients was prescribed at least one admission (OR: 3.7, 95% CI: 1.3–10.7, p = 0.02), and not working as a medication compared to the control group, although rates were high in health or social care worker (OR: 0.27, 95%: 0.08–0.9, p = 0.04). No both groups (76.7% vs 83.4% OR: 0.59, 95% CI: 0.39–0.89, p < 0.02). symptoms or other socio-demographic factors were predictive in the No information on medication was available for 28 mFND cases (8.7%) adjusted model. Univariate and adjusted analyses are presented in and 46 control group cases (7.1%). Table 2. We found no significant differences between proportions of mFND patients prescribed medication versus those prescribed no medications 3.3. Associations between patient characteristics and opioid analgesic, according to gender, marital status, pre-morbid employment status, co- antidepressant and antiepileptic prescriptions morbid physical illness, experience of childhood sexual or physical abuse, or patients with a carer versus those without. There were sig- Taking three commonly prescribed medications, we assessed char- nificant differences between mFND patients in employment, 19.4% of acteristics associated with prescriptions of opioid analgesics, anti- whom were prescribed medication, versus unemployed mFND patients, depressants and anti-epileptics, using mFND patients who weren't of whom 42.6% were prescribed medication (χ2 = 10.8, p = 0.001). prescribed these medications as a reference group. mFND patients who had a previous psychiatric inpatient admission In our univariate analysis the characteristics associated with opioid were more frequently prescribed medication compared to mFND pa- analgesic prescription were unemployment, receipt of welfare benefits, tients with no inpatient admission (40.5% versus 8.5%, χ2 = 17.6, lifetime use of walking aids, having a carer, functional weakness, back p = 0.001). mFND patients with a carer were more likely to have a pain, other pain, urinary and bowel dysfunction, and depression. medication prescription compared to patients with no carer (36.8% Adjusting for these in a logistic regression analysis we assessed the versus 21.3%, χ2 = 17.6, p < 0.001). functional motor symptoms associated with opioid analgesic prescrip- In mFND patients, the mean number of prescribed medications was tion. In this model, weakness (OR: 4.5, 95% CI: 1.7–11.7, p = 0.002) 4.77 (SD: 2.4), significantly higher than the 2.98 mean (SD: 2.7) in the and pain (excluding back pain) (OR: 4.5, 95% CI: 1.8–11.1, p = 0.001) control group (t (782) = 7.9, p = 0.001). The median prescription was were associated with opioid analgesic prescription. 4 (IQR: 5) in mFND, compared with 2 (IQR: 2) in the control group. In a univariate analysis, antidepressant prescriptions were asso- Fig. 2 (Supplementary material) displays a histogram showing the ciated with lifetime use of walking aids, and experiences of depression proportions of medications in both groups. Amongst mFND patients, the and anxiety. Accounting for these factors in an adjusted analysis of the most common prescriptions were antidepressants with 168 patients mFND symptoms linked to antidepressant use, we found having no (68%) receiving one or more, anti-epileptics (34% prescribed one or paralysis (OR: 0.45, 95% CI: 0.2–0.99, p = 0.05) and experiencing an- more) and non-opioid painkillers (32.8% prescribed one or more). xiety (OR: 2.9, 95% CI: 1.1–7.6, p = 0.03) were significantly predictive Opioid analgesics were the fourth most commonly prescribed medica- of antidepressant prescription. tion (31.2% of patients were prescribed one or more). Table 1 outlines A univariate analysis of factors found being male, British, married the proportion of medications in both groups. and experiencing childhood abuse were associated with anti-epileptic Compared to control patients, mFND patients had a higher like- medication. In an adjusted analysis we found no associations between lihood of being prescribed antidepressants, cardiac and anti- mFND symptoms and anti-epileptic prescriptions. The binary logistic hypertensive medications, statins, antihistamines (non-sedating), anti- regression results are outlined in Table 3. microbials, anti-asthma medications, corticosteroids, anti-epileptics, opioid analgesics, hormone replacement therapies (including medica- 4. Discussion tions for hormone imbalance and female contraception), proton-pump inhibitors, medications for bowel and urinary dysfunction, non-opioid The purpose of this study was to provide a cross-sectional descrip- analgesics, anti-inflammatories, anti-nauseas, nonsteroidal anti-in- tion of the medication patterns of mFND patients treated in a psy- flammatory drugs (NSAIDs) and muscle relaxants. chiatric setting, and the socio-demographic factors and mFND symp- Control patients were more likely to receive antipsychotic medica- toms associated with polypharmacy and certain prescriptions. tion and treatment for drug and alcohol addiction. There were no dif- Physicians prescribed a total of 1166 medications to 247 mFND ferences in prescriptions of beta blockers, central nervous system sti- patients. A slightly lower proportion of mFND patients received medi- mulants, nicotine replacement therapies, nutrients, hypnotics, cation compared to controls but rates in both groups were high. High benzodiazepines, anti-manic or anti-muscarinic medications. The odds rates in the psychiatric disorder control group are perhaps to be ex- ratios and confidence intervals of medication prescription in mFND pected but given the lack of evidence for the efficacy of medications in versus control group patients are displayed in Fig. 1. FND, the high prescription rate in the mFND group was surprising. Of patients with a prescription, mFND patients were given a higher pro- 3.2. Associations between mFND patient characteristics and polypharmacy portion and greater variety of medications. This heterogeneity has been reported previously in patients with multiple functional symptoms in Using the median number of medications, mFND patients were di- secondary care [12], and in fibromyalgia patients [43]. Polypharmacy vided into those prescribed five or more medications and those pre- in mFND patients was linked to co-morbid physical illness and previous scribed one to four. We examined the socio-demographic and symptom psychiatric inpatient admission. Rates of co-morbid physical illness in variables associated with prescription of five or more medications. FND is often high [29], and while it might partly account for poly- 108 (43.7%) mFND patients were prescribed five or more medica- pharmacy, when we assessed patients with no co-morbid organic con- tions, compared to 18.2% in the control group. In mFND only, un- ditions, mFND patients still had higher proportions of prescribed adjusted comparisons showed an association between prescription of medications, with a similar pattern in patients with no history of psy- five or more medications and older age, being older at the time of chiatric admission. symptom onset, receiving benefits, being unemployment, having pre- mFND patients were significantly more likely to have a prescription vious psychiatric inpatient admission, a physical health condition, a of one or more antihistamines, and the results indicated the drug was lifetime history of using a walking aid, experiencing childhood sexual more frequently prescribed to mFND patients as an anti-allergy and as a abuse, physical or sexual abuse in adulthood, and urinary or bowel sedative to control patients. Previous studies have reported higher rates dysfunction. of self-reported allergies in PNES patients compared to epileptic pa- In a binary logistic regression, there was a strong association tients [44], as well as heightened sensitivities and side effects to

97 N. O'Connell, et al. General Hospital Psychiatry 58 (2019) 94–102

Fig. 1. Forest plot showing odds of medication prescription in mFND patients compared to the control group. medications ([45], Nassim [46]). The presence of increased allergy and [50,51], where it may help shift patients' external locus of control to an non-specific medication side-effects in mFND deserves further ex- internal one, although deprescribing can be a difficult and complex amination and highlights the importance of a considered and colla- process [52]. borative prescription strategy [47]. Beyond psychological and physical health comorbidities, a variety of factors that we could not measure may contribute to these high rates 4.1. Antidepressants of somatic and psychotropic medication in FND. Diagnosing mFND can fi be complex, challenging and a time-consuming process for patients, We found a signi cantly higher proportion of mFND patients were particularly if presentation is atypical and there isn't access to suffi- prescribed at least one antidepressant (68%) compared to psychiatric fi ciently trained or experienced clinicians to consider FND, confirm it, controls, similar to previous ndings in patients with chronic functional and perhaps most importantly, explain it in a way that the patient symptoms in primary care [28]. This is perhaps surprising given that understands and accepts. During this process, patients likely accrue only 6.8% of mFND patients have a primary mood disorder diagnosis medications as they are referred within and between medical and and 5.4% have a secondary mood disorder diagnosis, compared to psychiatric specialities. One suggestion is that there is a dispropor- 22.7% and 25.1% with these diagnoses in the control group. In our tionate reliance on pharmacological solutions as clinicians often feel analysis of unstructured free text, we found 34% of mFND patients had fi obligated to try to do something to help a clearly distressed and dis- previous reports of low mood or depression, signi cantly lower than the abled patient, an option that might be preferred by both clinicians and 44.7% in the control group. There were anxiety rates of 17.4% and fi patients [48]. This argument gains credence in the context of over- 15.2% in the mFND and control groups respectively, with no signi cant ff stretched medical settings where medication prescriptions may be seen di erence. as a preferable passive form of treatment [49]. There have been recent Our rate is higher than previously reported rates, varying between recommendations for clinicians to utilise a reduction of medications 22 and 56%, in patients with unexplained visual loss, PNES patients, which lack a clear clinical indication in the therapeutic process itself somatoform disorder patients in primary care, mFND patients in a neurology and tertiary movement disorders clinic, and patients with

98 N. O'Connell, et al. General Hospital Psychiatry 58 (2019) 94–102

Table 2 Logistic regression model examining factors associated with polypharmacy in mFND patients (compared to prescription of between one and four medications).

Univariate 95% CI p value Adjusted 95% CI p value OR OR

Demographics Females versus males 1.60 0.9–2.9 0.12 2.1 0.7–6.2 0.19 British versus other ethnicity 1.2 0.7–2.1 0.43 1.0 0.4–2.7 0.99 Age at data collection 1.04 1.04–1.06 0.001 1.9 0.9–1.1 0.59 Age at symptom onset 1.02 1.0–1.04 0.04 1.0 0.9–1.1 0.09 Married versus unmarried 1.47 0.8–2.5 0.14 1.3 0.5–3.6 0.63 Welfare recipient versus receiving no beneﬁts 2.47 1.4–4.2 0.001 1.3 0.5–3.6 0.56 Employment Employed versus unemployed 0.27 0.13–0.57 0.001 0.31 0.08–1.2 0.09 Health or social care worker versus not employed in health or social care 0.88 0.45–1.7 0.72 0.27 0.08–0.9 0.04 Health Psychiatric inpatient stay versus no stay 2.33 1.4–3.9 0.001 3.5 1.2–10.2 0.02 Physical health condition versus no condition 3.18 1.6–6.1 0.001 7.5 2.1–26.7 0.002 Lifetime prevalence of walking aid versus no aid 2.0 1.2–3.5 0.01 0.7 0.3–1.7 0.40 Co-morbid non-motor functional symptoms versus no comorbid symptoms 1.15 0.7–1.9 0.60 0.7 0.3–1.8 0.43 Life events Childhood sexual abuse versus no CSA 2.92 1.4–5.9 0.003 1.6 0.5–5.5 0.43 Childhood physical abuse versus no CPA 1.74 0.9–3.4 0.10 1.8 0.5–6.7 0.39 Physical or sexual abuse in adulthood versus no abuse in adulthood 2.16 1.1–4.1 0.02 2.5 0.7–8.7 0.16 Carers Family carer versus not a family carer 0.64 0.3–1.6 0.34 0.48 0.1–2.3 0.37 Patient has a carer versus has no carer 0.59 0.3–1.0 0.06 2.1 0.8–5.2 0.13 Symptoms Tremor versus no tremor 1.2 0.7–1.9 0.55 1.9 0.6–5.9 0.25 Weakness versus no weakness 1.3 0.8–2.1 0.36 2.2 0.9–5.9 0.10 Back pain versus no back pain 1.0 0.5–2.1 0.91 1.8 0.4–7.3 0.39 Other pain types versus no other pain types 1.5 0.8–2.5 0.15 1.3 0.5–3.4 0.57 Numbness versus no numbness 1.1 0.6–1.9 0.79 2.5 0.9–7.1 0.08 Urinary/bowel dysfunction versus no urinary or bowel dysfunction 3.8 1.4–10.1 0.01 3.5 0.7–16.8 0.12 Paralysis versus no paralysis 0.8 0.4–1.7 0.62 0.44 0.13–1.4 0.18 Gait disorders versus no gait disorders 1.7 0.9–3.0 0.10 1.4 0.5–4.3 0.55 Depression present versus no depression present 1.2 0.7–1.9 0.51 1.4 0.5–3.5 0.52 Anxiety present versus no anxiety present 0.9 0.5–1.7 0.69 0.41 0.1–1.7 0.22 Fatigue present versus no fatigue present 0.8 0.3–2.1 0.66 1.9 0.3–13.0 0.51

CSA = childhood sexual abuse, CPA = childhood physical abuse. p values in bold denote statistically significant results. multiple functional symptoms in secondary care [12,53–57]. Our complaint in mFND, not just pain linked to the affected limb. higher rate partly reflects the specialist psychiatric setting from which Of patients on medication, 31% in our study received at least one the sample is drawn where patients will receive treatment for psy- opioid analgesic, a rate similar to reports in fibromyalgia [23] and chiatric co-morbidities, but compared to psychiatric patients from the unexplained gastrointestinal symptoms [24]. Like our antidepressant same Trust, antidepressant use is nonetheless high. results, it is difficult to ascertain whether patients are prescribed these The current guidance to prescribe only for co-morbidities in FND is drugs to treat co-morbidities, to treat mFND directly or to alleviate borne out in our adjusted findings as patients with a lifetime history of secondary pain arising from motor symptoms, like pain due to spasms, anxiety were nearly three times more likely to be prescribed anti- cramps and stiffness. Of note however, we did not find an association depressants. Antidepressants may have multiple actions, including between opioid analgesics and the presence of co-morbid physical possible central analgesic effects [58], and reductions in affective health conditions or co-morbid functional disorders like somatoform arousal and sleep dysfunction [59]. The relationship between anti- pain disorder, IBS or fibromyalgia. depressant treatment and mFND prognosis remains complicated by the In our univariate analysis, two associations with opioid analgesics presence of co-morbid depression and anxiety, as well as other factors emerged; severity (i.e. using a walking aid and having a carer) and such as attachment traits [60], emotional regulation [61], attentional socio-economic position (those unemployed and receiving welfare dysregulation [62], health anxiety and illness belief [63], and phobic benefits), similar to previous findings in patients with multiple func- avoidance [64]. As it stands, the effectiveness or otherwise of anti- tional symptoms and fibromyalgia [12,66]. Such associations may not depressants in this group and their mechanism of action remain poorly be specific to mFND as opioid prescription is higher in people with understood. mental health disorders [67] and in those with a lower socio-economic status in the general population [68,69]. A future study quantifying mFND severity in a more fine-grained 4.2. Opioid analgesics way would be helpful as previous research on somatisation reported that it was only patients with severe symptoms that engaged in opioid Opioid analgesics were more commonly prescribed in mFND pa- overuse [70]. Opioids were ranked as the most helpful medication tients than controls and the prescription was strongly associated with amongst fibromyalgia patients [71], although the evidence for their weakness, non-back related pain, and linked to urinary and bowel effectiveness is weak [72] and they have significant clinical risks, not dysfunction (although the association was not significant). While least pharmacological tolerance and dependence, bowel and bladder opioids are most commonly prescribed to treat pain, the associations dysfunction, cognitive impairment, hormonal changes, and even im- with urinary and bowel dysfunction may be due to the opioid medi- mune modulation. cation itself [65]. It is perhaps surprising that weakness, a lack of motor function, was associated with these drugs. This may be because pain inhibits patients from increasing physical activity as part of their re- 4.3. Anti-epileptics habilitation, leading to increased prescribing to relieve the impasse. However, it is likely the relationship is bidirectional with pain and lack 34% of mFND patients were prescribed one or more anti-epileptics. of activity reinforcing each other. In a case control study of functional Prescriptions were associated with being male, British, married and weakness Stone et al. [30] found widespread pain to be a common having experienced sexual abuse in childhood. Previous evidence on

99 .OCnel tal. et O'Connell, N.

Table 3 Logistic regression of functional symptoms that predict opioid analgesic, antidepressant and antiepileptic use in mFND.

Opioid analgesics Antidepressants Anti-epileptics

Univariate analysis Adjusted model Univariate analysis Adjusted model Univariate analysis Adjusted model

OR 95% CI p value OR 95% CI p value OR 95% CI p value OR 95% CI p value OR 95% CI p value OR 95% CI p value

Demographics Female versus male 1.1 0.6–2.1 0.70 0.9 0.5–1.7 0.80 0.43 0.2–0.8 0.005 British versus other ethnicity 1.1 0.6–1.9 0.79 0.81 0.4–1.4 0.45 1.95 1.1–3.4 0.02 Agea 1.0 0.9–1.0 0.94 0.99 0.9–1.0 0.54 0.98 0.9–1.0 0.06 Married versus single 0.74 0.4–1.3 0.27 1.6 0.9–2.7 0.10 1.4 0.8–2.3 0.25 Employment Employed versus unemployed 0.11 0.03–0.4 0.001 0.58 0.3–1.1 0.10 0.50 0.2–1.0 0.07 Welfare beneﬁts v no welfare 2.1 1.2–3.8 0.01 1.5 0.9–2.7 0.14 1.6 0.9–2.7 0.11 Health or social care worker versus not a health or social 0.65 0.3–1.4 0.28 0.8 0.4–1.7 0.63 0.6 0.3–1.3 0.20 care worker Illness Psychiatric inpatient stay versus no inpatient stay 1.6 0.9–2.7 0.10 1.4 0.8–2.4 0.27 0.83 0.5–1.4 0.49 Walking aid use versus no walking aid use 2.5 1.3–4.5 0.004 2.3 1.3–4.0 0.005 1.3 0.7–2.2 0.41 Physical illness versus no physical illness 0.84 0.5–1.6 0.58 0.98 0.5–1.8 0.94 1.1 0.6–2.0 0.81 Co-morbid non-motor functional symptoms versus no such 1.08 0.6–1.9 0.80 1.3 0.7–2.2 0.42 1.5 0.8–2.5 0.18 symptoms 100 Abuse Childhood sexual abuse versus no childhood sexual abuse 1.9 0.9–3.8 0.06 1.4 0.7–2.9 0.39 1.9 0.9–3.9 0.05 Childhood physical abuse versus no childhood physical 1.5 0.7–2.9 0.26 1.01 0.5–2.0 0.98 1.4 0.7–2.7 0.38 abuse Adult physical or sexual abuse versus no abuse in adulthood 0.85 0.4–1.7 0.64 0.96 0.5–1.9 0.91 0.9 0.4–1.7 0.63 Age of psychiatric symptom onseta 0.99 0.9–1.0 0.15 0.99 0.9–1.0 0.34 0.9 0.9–1.0 0.45 Caring Patient has a carer versus patient has no carer 2.3 1.3–4.1 0.01 1.4 0.8–2.5 0.26 0.96 0.5–1.7 0.89 Carer to family versus not a carer to family 0.7 0.3–2.1 0.54 2.03 0.7–6.3 0.22 0.62 0.2–1.8 0.37 Symptoms Tremor versus no tremor 0.9 0.5–1.6 0.79 0.65 0.20–1.9 0.42 0.9 0.5–1.6 0.69 0.80 0.4–1.7 0.54 0.85 0.5–1.5 0.57 1.9 0.9–4.2 0.10 Weakness versus no weakness 2.8 1.6–4.9 0.001 4.5 1.7–11.7 0.002 1.5 0.9–2.7 0.11 1.3 0.7–2.6 0.43 1.01 0.6–1.7 0.97 0.8 0.4–1.7 0.63 Back pain versus no back pain 2.6 1.3–5.2 0.01 0.46 0.2–1.4 0.16 1.0 0.5–2.2 0.94 0.98 0.4–2.4 0.96 1.34 0.7–2.7 0.42 1.1 0.5–2.9 0.77 Other pain types versus no other pain types 4.1 2.3–7.2 0.001 4.5 1.8–11.1 0.001 0.7 0.4–1.3 0.28 0.73 0.4–1.4 0.35 1.64 0.9–2.8 0.08 1.7 0.8–3.5 0.15 Numbness versus no numbness 1.1 0.6–2.0 0.76 0.90 0.4–2.3 0.83 1.06 0.6–1.9 0.86 0.9 0.4–1.9 0.81 1.4 0.8–2.5 0.29 1.1 0.5–2.2 0.85 Urinary/bowel dysfunction versus no urinary or bowel 4.5 1.8–11.2 0.001 4.5 0.9–21.4 0.06 1.02 0.4 –2.6 0.98 1.02 0.3–3.0 0.97 1.4 0.6–3.4 0.46 1.5 0.5–4.5 0.44 dysfunction Paralysis versus no paralysis 1.1 0.6–2.3 0.76 0.37 0.1–1.2 0.11 0.64 0.3–1.3 0.21 0.45 0.2–0.99 0.05 0.75 0.4–1.6 0.44 0.53 0.2–1.3 0.15 Gait disorders versus no gait disorders 0.9 0.5–1.8 0.86 0.7 0.3–2.1 0.55 0.90 0.5–1.7 0.74 0.68 0.3–1.4 0.30 0.82 0.4–1.6 0.54 0.73 0.3–1.6 0.43 Depression present versus no depression present 1.7 0.9–2.9 0.05 1.4 0.6–3.2 0.48 2.4 1.3–4.3 0.003 1.8 0.9–3.4 0.09 1.04 0.6–1.8 0.87 1.2 0.6–2.4 0.60 General HospitalPsychiatry58(2019)94– Anxiety present versus no anxiety present 0.84 0.4–1.7 0.62 0.54 0.15–1.9 0.36 2.6 1.2–5.9 0.02 2.9 1.1–7.6 0.03 1.2 0.6–2.3 0.61 0.9 0.4–2.3 0.83 Fatigue present versus no fatigue present 1.1 0.4–3.1 0.85 1.4 0.2–9.3 0.70 0.73 0.3–1.9 0.52 0.7 0.2–2.2 0.53 0.98 0.4–2.7 0.97 1.01 0.3–3.6 0.98

Adjusted model accounts for factors significant in the univariate analysis. p values in bold denote statistically significant results. a For a 1-year increase in age, and one additional functional symptom. 102 N. O'Connell, et al. General Hospital Psychiatry 58 (2019) 94–102 anti-epileptics in mFND found 27% of functional movement disorder 4.5. Conclusions patients were prescribed anti-epileptics [9], and anti-epileptics were commonly prescribed in children with mFND [73]. This is the first study to describe medication patterns in mFND Anti-epileptics' multiple indications makes it difficult to draw con- patients. These patients are prescribed an extensive range of psychiatric clusions on their usage. A previous study reported that anti-epileptics and physical health medication, most commonly anti-depressants, anti- were prescribed by psychiatrists most frequently to treat bipolar dis- epileptics and analgesics. This polypharmacy may be partially ex- order and anxiety, while neurologists tended to prescribe them to treat plained by higher rates of physical co-morbidities but may also reflect seizures, migraine and pain [74]. In our sample, anti-epileptics may functional symptoms in other bodily systems, higher symptom re- have been prescribed to modify pain and reduce anxiety [75] and to porting combined with a lack of therapeutic options for clinicians manage co-morbid seizures (epileptic and or undiagnosed PNES), al- managing patients with complex functional and ‘organic’ conditions though no specific symptoms associations or links with co-morbid and chronic pain. As there are no evidence-based pharmacological physical illness were observed. In an additional analysis, not reported therapies, it is important to carefully assess the benefit to risk ratio of above, we found no significant difference in the rates of anti-epileptic prescribing in this complex and poorly understood disorder. prescriptions between mFND patients with co-morbid non-epileptic seizures and those without. Another possible explanation for their high Acknowledgements prescription rate may be that the motor symptoms are themselves a result of the administration of anti-epileptics [76], although this is We thank the staff at the BRC Nucleus for their advice and expertise, unlikely in this tertiary psychiatric setting where the functional diag- in particular Megan Pritchard, Amelia Jewel and Debbie Cummings. nosis is often well-established. Declaration of interests

4.4. Strengths and limitations The authors declare no conﬂicting interests.

This study is the first to describe detailed medication usage in a Funding large sample of mFND patients. The use of routine clinical records al- lows for a naturalistic study without the selection biases common to NOC was funded by the National Institute for Health Research trials or studies requiring patient recruitment, increasing the re- (NIHR) Biomedical Research Centre at South London and Maudsley presentativeness of our results. Access to anonymous health records NHS Foundation Trust and King’s College London. TN is funded by an allowed us to compile detailed socio-demographic and health in- NIHR Clinician Scientist Award. ASD is supported by the UCLH formation on mFND patients, and the use of a contemporaneous control National Institute of Health Research (NIHR) Biomedical Research group enabled us to report the specific risk of medication prescriptions Centre. The views expressed are those of the authors and not necessarily in mFND patients. those of the NHS, the NIHR or the Department of Health and Social The use of electronic health records brings certain limitations. While Care. we can assess patterns in prescriptions, we cannot report their adherence, effectiveness or side effects. For drugs with multiple clinical Appendix A. Supplementary data indications, such as anti-epileptics and antidepressants, we do not know the primary reason for their prescription or where and by whom Supplementary data to this article can be found online at https:// medication was initiated. For each patient, we recorded con- doi.org/10.1016/j.genhosppsych.2019.04.004. temporaneous medication information from their clinical notes, so we cannot report treatment duration or discontinuation. In addition, we References cannot account for any data entry errors made in health records, and it is possible that there may be under-reporting of medications, particu- [1] American Psychiatric Association. Diagnostic and statistical manual of mental dis- larly as we capped the number of medications at 12, although this is orders. Washington DC: American Psychiatric Association; 2013. [2] Carson A, Stone J, Hibberd C, Murray G, Duncan R, Coleman R, et al. Disability, unlikely to cause systematic bias. Our method will only capture self- distress and unemployment in neurology outpatients with symptoms ‘unexplained prescriptions if this has been disclosed to the clinician. Our control by organic disease’. J Neurol Neurosurg Psychiatry 2011;82:810–3. [3] Anderson KE, Gruber-Baldini AL, Vaughan CG, Reich SG, Fishman PS, Weiner WJ, group included patients with psychotic spectrum disorders and addic- et al. Impact of psychogenic movement disorders versus Parkinson's on disability, tions which will likely accentuate differences in antipsychotic medica- quality of life, and psychopathology. Mov Disord 2007;22:2204–9. tion and drug and alcohol treatment risks however overcontrolling for [4] Vroegop S, Dijkgraaf MGW, Vermeulen M. Impact of symptoms in patients with fi functional neurological symptoms on activities of daily living and health related such conditions would have obscured other interesting ndings in the quality of life. J Neurol Neurosurg Psychiatry 2013;84:707–8. contrast of FND and other psychiatric patients. This retrospective study [5] Gelauff J, Stone J, Edwards M, Carson A. The prognosis of functional (psychogenic) included patients treated over a ten-year period so it is possible more motor symptoms: a systematic review. J Neurol Neurosurg Psychiatry 2014;85:220–6. recently treated patients will have been prescribed drugs not developed [6] Espay AJ, Aybek S, Carson A, Edwards MJ, Goldstein LH, Hallett M, et al. Current in 2006. In addition, prescribing patterns themselves may have shifted concepts in diagnosis and treatment of functional neurological disorders. JAMA – fi Neurol 2018;75:1132 41. due to changes in DSM-5 criteria, speci cally the removal of the need to [7] Voon V, Lang AE. Antidepressant treatment outcomes of psychogenic movement prove a temporal link between psychological factors and the onset of disorder. J Clin Psychiatry 2005;66:1529–34. ff symptoms, and advances in the education and training of new clin- [8] Rampello L, Ra aele R, Nicoletti G, Le Pira F, Malaguarnera M, Drago F. Hysterical neurosis of the conversion type: therapeutic activity of neuroleptics with different icians, something we are unable to account for in our analysis. We es- hyperprolactinemic potency. Neuropsychobiology 1996;33:186–8. tablished the drug categorisations based on the collected data and [9] Ganos C, Aguirregomozcorta M, Batla A, Stamelou M, Schwingenschuh P, Münchau A, et al. Psychogenic paroxysmal movement disorders – clinical features and diag- therapeutic usage but given the multi-action nature of certain drugs and nostic clues. Parkinsonism Relat Disord 2014;20:41–6. the many sub-classifications, it is possible that our definitions and use [10] Lafrance Jr. WC, Keitner GI, Papandonatos GD, Blum AS, Machan JT, Ryan CE, et al. of a miscellaneous category missed some meaningful associations. Pilot pharmacologic randomized controlled trial for psychogenic nonepileptic seizures. Neurology 2010;75:1166–73. Finally, this study was set in a large psychiatric NHS Trust. The Trust [11] Pintor L, Bailles E, Matrai S, Carreno M, Donaire A, Boget T, et al. Efficiency of receives referrals as a tertiary neuropsychiatry service, which limits our venlafaxine in patients with psychogenic nonepileptic seizures and anxiety and/or – ability to generalise to services without neuropsychiatry input or to depressive disorders. J Neuropsychiatry Clin Neurosci 2010;22:401 8. [12] Agger JL, Fink PK, Gormsen LK, Jensen JS, Schröder A. The use of prescription other healthcare systems. medication in 239 patients with multiple functional somatic syndromes. Gen Hosp Psychiatry 2018;51:96–105.

101 N. O'Connell, et al. General Hospital Psychiatry 58 (2019) 94–102

[13] Agger JL, Schröder A, Gormsen LK, Jensen JS, Jensen TS, Fink PK. Imipramine [43] Robinson RL, Kroenke K, Mease P, Williams DA, Chen Y, D'souza D, et al. Burden of versus placebo for multiple functional somatic syndromes (STreSS-3): a double- illness and treatment patterns for patients with fibromyalgia. Pain Med blind, randomised study. Lancet Psychiatry 2017;4:378–88. 2012;13:1366–76. [14] Kleinstäuber M, Witthöft M, Steffanowski A, Van Marwijk H, Hiller W, Lambert MJ. [44] Robbins NM, Larimer P, Bourgeois JA, Lowenstein DH. Number of patient-reported Pharmacological interventions for somatoform disorders in adults. Cochrane allergies helps distinguish epilepsy from psychogenic nonepileptic seizures. Database Syst Rev 2014;11:CD010628. Epilepsy Behav 2016;55:174–7. [15] Ford AC, Talley NJ, Schoenfeld PS, Quigley EM, Moayyedi P. Efficacy of anti- [45] Park JH, Bokma J, Chapple K, Caplan JP. A retrospective study of polyallergy as a depressants and psychological therapies in irritable bowel syndrome: systematic marker of nonepileptic seizures in the epilepsy monitoring unit. Psychosomatics review and meta-analysis. Gut 2009;58:367–78. 2014;55:566–71. [16] Ruepert L, Quartero AO, De Wit NJ, Van Der Heijden GJ, Rubin G, Muris JWM. [46] Matin Nassim, Young Sigrid S, Williams Benjamin, Lafrance Jr. W Curt, King Julie Bulking agents, antispasmodics and antidepressants for the treatment of irritable N, Caplan David, et al. Neuropsychiatric associations with gender, illness duration, bowel syndrome. Cochrane Database Syst Rev 2011;8:CD003460. work disability, and motor subtype in a U.S. functional neurological disorders clinic [17] Kroenke K, Krebs EE, Bair MJ. Pharmacotherapy of chronic pain: a synthesis of population. J Neuropsychiatry Clin Neurosci 2017;29:375–82. recommendations from systematic reviews. Gen Hosp Psychiatry 2009;31:206–19. [47] Barsky AJ, Saintfort R, Rogers MP, Borus JF. Nonspecific medication side effects [18] Fishbain DA, Cutler RB, Rosomoff HL, Rosomoff RS. Do antidepressants have an and the nocebo phenomenon. JAMA 2002;287:622–7. analgesic effect in psychogenic pain and somatoform pain disorder? A meta-ana- [48] Jackson JL, Kroenke K. Managing somatization: medically unexplained should not lysis. Psychosom Med 1998;60:503–9. mean medically ignored. J Gen Intern Med 2006;21:797–9. [19] O'Malley PG, Jackson JL, Santoro J, Tomkins G, Balden E, Kroenke K. [49] Netherlands Institute of Mental Health and Addiction. Multidisciplinary guideline Antidepressant therapy for unexplained symptoms and symptom syndromes. J Fam for medically unexplained symptoms and somatoform disorders. Houten: Ladenius Pract 1999;48:980–90. Communicatie BV; 2010. [20] Wessely S. Review: antidepressants reduce pain intensity in psychogenic pain or [50] McCormack R, Moriarty J, Mellers J, Shotbolt P, Pastena R, Landes N, et al. somatoform pain disorder. Evid Based Ment Health 1999;2:13. Specialist inpatient treatment for severe motor conversion disorder: a retrospective [21] Heintz CE, Van Tricht MJ, Van Der Salm SM, Van Rootselaar AF, Cath D, Schmand comparative study. J Neurol Neurosurg Psychiatry 2014;85:895–900. B, et al. Neuropsychological profile of psychogenic jerky movement disorders: im- [51] Stone J. Functional neurological disorders: the neurological assessment as treat- portance of evaluating non-credible cognitive performance and psychopathology. J ment. Pract Neurol 2016;16:7–17. Neurol Neurosurg Psychiatry 2013;84:862–7. [52] Salisbury H. Helen Salisbury: the perils of deprescribing. BMJ 2019;364:l666. [22] Kouyanou K, Pither CE, Rabe-Hesketh S, Wessely S. A comparative study of ia- [53] de Waal MWM, Arnold IA, Eekhof JAH, Assendelft WJJ, Van Hemert AM. Follow-up trogenesis, medication abuse, and psychiatric morbidity in chronic pain patients study on health care use of patients with somatoform, anxiety and depressive dis- with and without medically explained symptoms. Pain 1998;76:417–26. orders in primary care. BMC Fam Pract 2008;9:5. [23] Fitzcharles M-A, Ste-Marie PA, Gamsa A, Ware MA, Shir Y. Opioid use, misuse, and [54] Duncan R, Graham CD, Oto M, Russell A, Mckernan L, Copstick S. Primary and abuse in patients labeled as fibromyalgia. Am J Med 2011;124:955–60. secondary care attendance, anticonvulsant and antidepressant use and psychiatric [24] Sayuk GS, Kanuri N, Gyawali CP, Gott BM, Nix BD, Rosenheck RA. Opioid medi- contact 5–10 years after diagnosis in 188 patients with psychogenic non-epileptic cation use in patients with gastrointestinal diagnoses vs unexplained gastro- seizures. J Neurol Neurosurg Psychiatry 2014;85:954–8. intestinal symptoms in the US Veterans Health Administration. Aliment Pharmacol [55] Kranick S, Ekanayake V, Martinez V, Ameli R, Hallett M, Voon V. Psychopathology Ther 2018;47:784–91. and psychogenic movement disorders. Mov Disord 2011;26:1844–50. [25] Vercoulen JH, Swanink CM, Fennis JF, Galama JM, Van Der Meer JW, Bleijenberg [56] Noyes R, Happel RL, Muller BA, Holt CS, Kathol RG, Sieren LR, et al. Fluvoxamine G. Prognosis in chronic fatigue syndrome: a prospective study on the natural course. for somatoform disorders: an open trial. Gen Hosp Psychiatry 1998;20:339–44. J Neurol Neurosurg Psychiatry 1996;60:489–94. [57] O'Leary ÉD, Mcneillis B, Aybek S, Riordan-Eva P, David AS. Medically unexplained [26] Gazzola DM, Carlson C, Rugino A, Hirsch S, Starner K, Devinsky O. Psychogenic visual loss in a specialist clinic: a retrospective case–control comparison. J Neurol nonepileptic seizures and chronic pain: a retrospective case-controlled study. Sci 2016;361:272–6. Epilepsy Behav 2012;25:662–5. [58] Lynch ME. Antidepressants as analgesics: a review of randomized controlled trials. J [27] Hantke NC, Doherty MJ, Haltiner AM. Medication use profiles in patients with Psychiatry Neurosci 2001;26:30–6. psychogenic nonepileptic seizures. Epilepsy Behav 2007;10:333–5. [59] Clouse RE, Lustman PJ, Geisman RA, Alpers DH. Antidepressant therapy in 138 [28] olde Hartman TC, Lucassen PLBJ, Van De Lisdonk EH, Bor HHJ, Van Weel C. patients with irritable bowel syndrome: a five-year clinical experience. Aliment Chronic functional somatic symptoms: a single syndrome? Br J Gen Pract Pharmacol Ther 1994;8:409–16. 2004;54:922–7. [60] Jalilianhasanpour R, Ospina JP, Williams B, Mello J, Maclean J, Ranford J, et al. [29] Pareés I, Kojovic M, Pires C, Rubio-Agusti I, Saifee TA, Sadnicka A, et al. Physical Secure attachment and depression predict 6-month outcome in motor functional precipitating factors in functional movement disorders. J Neurol Sci neurological disorders: a prospective pilot study. Psychosomatics 2018. (Aug 29. 2014;338:174–7. pii: S0033-3182(18)30422-5). [30] Stone J, Warlow C, Sharpe M. The symptom of functional weakness: a controlled [61] Sojka P, Bares M, Kasparek T, Svetlak M. Processing of emotion in functional study of 107 patients. Brain 2010;133:1537–51. neurological disorder. Front Psych 2018;9:479. [31] Alexander RW, Aaron LA, Alberts KR, Martin MY, Stewart KE, Bradley LA, et al. [62] Edwards MJ, Adams RA, Brown H, Parees I, Friston KJ. A Bayesian account of Sexual and physical abuse in women with fibromyalgia: association with outpatient ‘hysteria’. Brain 2012;135:3495–512. health care utilization and pain medication usage. Arthritis Rheum [63] Bakvis P, Roelofs K, Kuyk J, Edelbroek PM, Swinkels WA, Spinhoven P. Trauma, 1998;11:102–15. stress, and preconscious threat processing in patients with psychogenic nonepileptic [32] Olde Hartman TC, Hassink-Franke LJ, Lucassen PL, Van Spaendonck KP, Van Weel seizures. Epilepsia 2009;50:1001–11. C. Explanation and relations. How do general practitioners deal with patients with [64] Stone J, Gelauff J, Carson A. A “twist in the tale”: altered perception of ankle po- persistent medically unexplained symptoms: a focus group study. BMC Fam Pract sition in psychogenic dystonia. Mov Disord 2012;27:585–6. 2009;10:68. [65] Panchal SJ, Müller-Schwefe P, Wurzelmann JI. Opioid-induced bowel dysfunction: [33] Richardson RD, Engel Jr. CC, Mcfall M, Mcknight K, Boehnlein JK, Hunt SC. prevalence, pathophysiology and burden. Int J Clin Pract 2007;61:1181–7. Clinician attributions for symptoms and treatment of gulf war–related health con- [66] Fitzcharles M-A, Faregh N, Ste-Marie PA, Shir Y. Opioid use in fibromyalgia is as- cerns. Arch Intern Med 2001;161:1289–94. sociated with negative health related measures in a prospective cohort study. Pain [34] Rommelfanger KS. Attitudes on mind over matter: physician views on the role of Res Treat 2013;2013:7. placebo in psychogenic disorders. AJOB Neurosci 2013;4:9–15. [67] Sullivan MD, Edlund MJ, Zhang L, Unützer J, Wells KB. Association between mental [35] Shamy MCF. The treatment of psychogenic movement disorders with suggestion is health disorders, problem drug use, and regular prescription opioid use. Arch Intern ethically justified. Mov Disord 2010;25:260–4. Med 2006;166:2087–93. [36] Barsky AJ, Orav E, Bates DW. Somatization increases medical utilization and costs [68] Ruscitto A, Smith BH, Guthrie B. Changes in opioid and other analgesic use independent of psychiatric and medical comorbidity. Arch Gen Psychiatry 1995–2010: repeated cross-sectional analysis of dispensed prescribing for a large 2005;62:903–10. geographical population in Scotland. Eur J Pain 2015;19:59–66. [37] McGorm K, Burton C, Weller D, Murray G, Sharpe M. Patients repeatedly referred to [69] Zhou C, Yu NN, Losby JL. The association between local economic conditions and secondary care with symptoms unexplained by organic disease: prevalence, char- opioid prescriptions among disabled Medicare beneficiaries. Med Care acteristics and referral pattern. Fam Pract 2010;27:479–86. 2018;56:62–8. [38] Perera G, Broadbent M, Callard F, Chang C, Downs J, Dutta R, et al. Cohort profile [70] Trafton JA, Cucciare MA, Lewis E, Oser M. Somatization is associated with non- of the South London and Maudsley NHS Foundation Trust Biomedical Research adherence to opioid prescriptions. J Pain 2011;12:573–80. Centre (SLaM BRC) Case Register: current status and recent enhancement of an [71] Bennett RM, Jones J, Turk DC, Russell IJ, Matallana L. An internet survey of 2,596 Electronic Mental Health Record-derived data resource. Br Med J 2016;6. people with fibromyalgia. BMC Musculoskelet Disord 2007;8:27. [39] Joint Formulary Committee. BNF 74: September 2017. London: Pharmaceutical [72] Clauw DJ. Fibromyalgia: a clinical review. Jama 2014;311:1547–55. Press; 2017. [73] Ferrara J, Jankovic J. Psychogenic movement disorders in children. Mov Disord [40] Haider SI, Johnell K, Weitoft GR, Thorslund M, Fastbom J. The influence of edu- 2008;23:1875–81. cational level on polypharmacy and inappropriate drug use: a register-based study [74] Cascade E, Kalali AH, Weisler RH. Varying uses of anticonvulsant medications. of more than 600,000 older people. J Am Geriatr Soc 2009;57:62–9. Psychiatry (Edgmont) 2008;5:31–3. [41] Hovstadius B, Hovstadius K, Åstrand B, Petersson G. Increasing polypharmacy - an [75] Mula M, Pini S, Cassano GB. The role of anticonvulsant drugs in anxiety disorders: a individual-based study of the Swedish population 2005–2008. BMC Clin Pharmacol critical review of the evidence. J Clin Psychopharmacol 2007;27:263–72. 2010;10:16. [76] Zadikoff C, Munhoz RP, Asante AN, Politzer N, Wennberg R, Carlen P, et al. [42] O'Connell N, Nicholson TR, Wessely S, David AS. Characteristics of patients with Movement disorders in patients taking anticonvulsants. J Neurol Neurosurg motor functional neurological disorder in a large UK mental health service: a case- Psychiatry 2007;78:147–51. control study. Psychol Med 2019:1–10.

102